Lucenesearch-RebuildIndex-Job and Lucenesearch-UpdateIndex-Job run for long time

[Former Member]

Good evening, we're working on hybris 6.3.0.0 (ex- novo installation on a new machine) with some customization on solr search providers and facets (hoping this details could help i the resolution of the issue).

During the monitoring phase of the project, we have been noticing that LuceneIndex Cronjobs ("Lucenesearch-RebuildIndex-Job" and "Lucenesearch-UpdateIndex-Job") run for very long time (in terms of hours or days sometimes) decreasing obviously the system performances.

Following our questions: 1) Is there a way to understand if these jobs can be deactivated? 2) Which is their purpose? 3) Could they affect /being affected solr search or indexing? 4) Why do they need all this time?

In the application log, we've found also:

03.02.2018 12:15:31: ERROR: Caught throwable error while accessing lucenesearch index hmc de.hybris.platform.lucenesearch.jalo.IndexAccessException: error while accessing lucenesearch index hmc[HY-0] at de.hybris.platform.lucenesearch.jalo.LuceneIndex.removeItemFromIndexImpl(LuceneIndex.java:875) at de.hybris.platform.lucenesearch.jalo.LuceneIndex.removeItemFromIndex(LuceneIndex.java:818) at de.hybris.platform.lucenesearch.jalo.UpdateIndexJob.performCronJob(UpdateIndexJob.java:335) at de.hybris.platform.cronjob.jalo.Job.execute(Job.java:1358) at de.hybris.platform.cronjob.jalo.Job.performImpl(Job.java:814) at de.hybris.platform.cronjob.jalo.Job.access$1(Job.java:767) at de.hybris.platform.cronjob.jalo.Job$JobRunable.run(Job.java:686) at de.hybris.platform.util.threadpool.PoolableThread.run(PoolableThread.java:195)

Searching on this portale we've found only this thread: https://answers.sap.com/questions/12766074/luceneindex-jobs-slowing-down-hybris.html

but was not helpful in the issue analysis.

Thanks in advance for the support.

[Former Member]

Hi Marcello,

Yes - these jobs can be deactivated, either by impex (removing the jobs) or manually be deleting. Their purpose is to index search data within the HMC.

If there has been a lot of data updates - ie data imported, naturally occuring due to a new project and these jobs have not been ran regularly, then they can take a long time to complete and affect performance of other running jobs, for example the front end solr index or other running cronjobs for that matter.

I can't precisely pinpoint what is causing your specific exception, without more information but given that this should be disabled/removed, this will also negate the exception investigation.